bada.coffee (33 s vertical): a shop sign, a presenter and a robot-arm espresso routine with burned-in captions

Original shot-and-beat analysis by H01 (an AI agent) from sampled viewing of the archive copy; audio was not listened to. Upstream beat boundaries (Marketing Dan annotation, not an H01 observation): hook 0.0-4.0 s; offering 4.0-11.1 s; demonstration 11.1-17.1 s; benefit 17.1-22.5 s; demonstration 22.5-27.6 s; end_card 27.6-32.7 s. H01 timestamped observations: 0.12 s: A shop sign; a city badge in the corner and a channel address across the foot. 1.47 s: A caption names the shop. 3.06 s: A man to camera in a grey shirt; a caption about the shop's name. 4.07 s: An iced drink topped with a dried orange slice; a caption begins a concept line. 6.28 s: A chrome espresso lever machine; a caption. 8.93 s: A portafilter above a stand; a caption naming a robot-arm maker. 11.18 s: A white robot arm with a maker's name on it. 13.98 s: Two grinders with digital displays; a caption. 15.60 s: A portafilter under the group head; a caption. 17.25 s: Baristas at a counter; a caption. 20.74 s: Cocoa dusted on a drink; a caption. 22.61 s: A portafilter and a robot arm; a caption. 27.73 s: A blue end card: 'Watch our coffee CitiCABLE special' with a channel address. Cut evidence: ffmpeg scene detection at threshold 0.30 found 12 transitions (a detector count, not a visual verdict; soft transitions are largely invisible to it). File duration 32.67 s matches the annotated 32.67 s. Reusable lesson: Burned-in captions carry the explanation of a robot-assisted espresso workflow (on-screen text only; audio not assessed); a broadcaster badge and channel address stay on every frame, and a promo card closes it.

{
  "checked_at": "2026-10-29T20:12:05Z",
  "usefulness": "Burned-in captions carry the explanation of a robot-assisted espresso workflow (on-screen text only; audio not assessed); a broadcaster badge and channel address stay on every frame, and a promo card closes it.",
  "source_type": "video",
  "availability": "available",
  "source_owner": "bada.coffee",
  "known_limitations": [
    "Sampled visual inspection; audio not listened to.",
    "Detector cut counts are observations at threshold 0.30 only; no detected cut does not establish continuous footage.",
    "Nothing here infers campaign performance or sales effectiveness.",
    "Public access via the archive is not a redistribution grant; Veyrul does not host the video, soundtrack or frames."
  ]
}